DarkNeurons: An Automated Deep Learning ...

DarkNeurons: An Automated Deep Learning Library

Oct 26, 2020

Hello to all AI Enthusiasts out there, Hope you are doing good in this pandemic era and improving your skills.

DarkNeurons is an Open-Source implementation of Automatic Deep Learning Library which can reduce the time and Complexity for non-technical users to train their own networks without Compromising Accuracies for Classification of Images and Object Detection, the most demanding techniques for Autonomous Systems and Medical Fields.

" By augmenting human performance, AI has the potential to markedly improve productivity, efficiency, workflow, accuracy and speed, both for physicians and for patients … What I’m most excited about is using the future to bring back the past: to restore the care in healthcare. " - Eric Topol


  • Installation

  • Classification

  • Object Detection


Two ways to install Library:

$ pip install DarkNeurons


git clone https://github.com/Tushar-ml/DarkNeuron.git

cd DarkNeuron

python setup.py install

Providing the Summary of Methods in DarkNeurons below, Detailed Instructions can be seen over here: Show Your Support by Forking the Repository and by Providing contribution to it.



DarkNeuron Classification has the feature of implementing Pretrained Models on ImageNet Data. Users can directly train pre-trained models or can retrain their own models. Models provided are:

  • InceptionV3

  • Xception

  • ResNet50

  • VGG16

  • VGG19 Further will be added on upcoming releases.

  • Initialization:

from DarkNeurons import Classify_Images

classify = Classify_Images( working directory )

  • Preparation and Preprocessing of Data:

    Train , Valid , Test = classify.Preprocess_the_Image( )

  • Model Creation

    model = classify.Create_the_Model( )

  • Model Training

    model = classify.Train_the_Model( )

  • Prediction

    classify.Predict_from_the_Model( )

  • Visualization of Metrics and Prediction

    classify.Visualize_the_Metrics( )

    classify.Visualize_the_Predictions( )


Object Detection Model uses the YOLOv4 version for Detection. It is the fastest algorithm known for Detection

  • Initialization:

    from DarkNeurons import YOLOv4

    yolo = YOLOv4( working directory )

  • Preparation of Data:

    yolo.Prepare_the_Data( )

  • Model Training

    yolo.Train_the_Yolo( )

  • Detection

    yolo.Detect( )

In the Detection part, you can detect from Image, Video or even Live Capturing from Screen Detection.

Further Releases:

  • Visualization of Artificial Neural Networks

  • Hybridization or Fusion of Two Different Models

  • Object Tracking using DeepSORT

Improve your Deep Learning Skills by Contributing to the Open-Source Libraries.

Keep Coding, Keep Growing with AI

For Further Open-Source Libraries and Tutorials, Follow me for the newsletters directly to your emails. For Contributing topics for this post, connect with me on:

Thank You!!!!

Enjoy this post?

Buy tushargai a coffee